Technology advances often overshadow persistent issues. While new AI models draw attention, large social platforms used by hundreds of millions of teenagers continue to host harmful and explicit material during key developmental years.
Meta operates Facebook, Instagram, WhatsApp and Messenger, reaching about 3.5 billion daily users. The company has integrated its own AI systems and recommendation tools over recent years, though outcomes have not fully met expectations.
Meta developed PyTorch, an open-source framework now common in production, and released the Llama model family. In June 2025, it acquired a 49 percent stake in Scale AI for 14.3 billion dollars. Leadership changes followed, shifting focus to internal product applications.
Meta released its Muse Image model in July 2025. A feature allowing image generation from public profiles was enabled by default and later disabled after concerns arose.
Earlier scrutiny in 2021 followed whistleblower disclosures about engagement-driven ranking systems that can amplify risky content for vulnerable users. Instagram continued expanding algorithmic feeds and short-form video despite these warnings.
Roughly 100 million photos and videos are uploaded daily to Instagram. Meta reported under one percent of posts violating rules in late 2025, while filing over two million reports on suspected exploitation, many involving child sexual abuse material.
A BBC investigation identified paid ads in India promoting such material via Telegram. Test accounts received sexualized recommendations within days. Meta initially deemed some ads compliant before removing them after further review. Indian authorities requested explanations.
